Starte nie bei null

Daily scan of all synced creator content: captions are keyword-matched, unmatched stories are transcribed by AI (speech + on-screen text) and checked for your brand terms. Hits get marked on each creator's detail page — so you see who actually posted about you.

Every morning: check each creator's story quota — remind the ones falling behind shortly before month end, congratulate the ones who hit their goal. Deduped per month, so it never double-sends.

The moment a customer completes a survey, mint a one-time Shopify code for a free product (one code per customer, quantity-safe fixed amount) and email it to them automatically. Pair it with an email campaign whose {{hosted_link}} sends each customer their personal survey page.

Every Monday 09:00, take the top 3 active ads by spend (ROAS above 2 over 7 days) and generate 3 similar marketing angles each — saved into a Canvas list.

Every morning at 07:00: active ads (campaigns on Google — no per-ad tracking there) that spent over 100 in 7 days at ROAS below 1 get a short funnel diagnosis — which stage leaks and the estimated weekly bleed — plus a queued pause and an email that tells you WHY it burns.

Every morning at 07:00: ad sets running ROAS above 3 on meaningful spend — and stable or improving vs the previous week — get a scaling-headroom analysis (frequency saturation, CPM pressure, CTR freshness), a 20% budget increase queued for approval and the full numbers in your inbox.

Every morning at 08:00: campaigns whose CPA jumped more than 30% vs the previous 3 days (on real spend) get a root-cause decomposition — CPA moves through CPM, CTR or CR, so the analysis names WHICH factor moved, rules out what the numbers contradict, and gives the fix.

Every Monday at 09:00: every campaign above the spend floor gets classified from a 14-vs-14-day comparison — compounding winner, stable, slow bleeder (the expensive kind nobody notices) or burner — with the projected monthly cost of each leak and one action per campaign.

Every 6 hours, scan new creator stories, AI-label each one (hook, product, emotion, format), then generate 3 fresh marketing angles per story — saved into a Canvas list, ready to brief.

Every Monday, a per-creator content scorecard from the AI labels of their stories and posts — strengths, weaknesses and quality assessment for every creator you work with.

Tie creators to real money — the commission engine's attributed revenue and conversions per creator, with a weekly scale / observe / drop readout.

Every Monday 09:00, take your creatives labeled Winner, let AI extract what makes their hooks work and generate 5 fresh hook variations per winner — saved into a Canvas list.

Every Monday 09:00, AI researches hook patterns currently trending on Meta and TikTok in your industry and proposes 10 fresh hooks adapted to your brand — saved into a Canvas list and emailed to you.

The moment a creative flips to Fatigued, AI generates 5 hook swaps that keep the proven body but replace the worn-out opening — saved into a Canvas list, ready to brief.

On the 1st of every month, find paused ads that had ROAS above 2 over the last 90 days and duplicate them into a campaign you pick — created paused and queued for approval.

Every morning at 07:00, find active ad sets that spent over 50 in the last 3 days with ROAS below 1 and cut their budget by 20% — queued for approval, with a 3-day cooldown per ad set.

Every morning at 07:00, find active ad sets with ROAS above 3 over the last 3 days and raise their budget by 20% — queued for approval, with a 3-day cooldown per ad set.

The moment a creative flips to Winner, AI dissects it into its elements — hook, visual, copy, CTA, offer — explains which one carries the performance, and saves the breakdown as a briefing in a Canvas list.

A creative flips to Winner, holds its ROAS above 2 for 3 more days — then AI prepares the next iteration round: 5 variations that each change exactly ONE element (hook, visual, copy, CTA or offer), saved into a Canvas list.

Every morning at 08:00, check creatives in Testing: anything that spent over 50 in the last 7 days with ROAS above 3 gets promoted to Winner — which in turn fires your winner workflows (dissection, iteration batch).

Every morning at 08:00, check creatives in Testing: anything that spent over 50 in the last 7 days with ROAS below 1 gets labeled Flop and a pause is queued for approval — the iteration slot is freed for the next variation.

The moment a creative flips to Flop, AI compares it against your current winners, explains what most likely killed it (hook, offer, audience fit, format) and saves the learning to a Canvas list — so the next iteration round doesn't repeat the mistake.

A creative flips to Fatigued → let it rest for 45 days, then queue a reactivation for approval — often the cheapest 'new' ad you can launch, the audience has forgotten it.

On the 1st of each quarter, dig through paused ads that hit ROAS above 2 at any point in the last 365 days, and duplicate the best 10 into a campaign you pick — last year's seasonal winners come back exactly when their season returns.

The moment an ad flips to Winner, duplicate it into a second campaign you pick (scaling, retargeting or another audience) — created paused and queued for approval.

On the 1st of every month, take your creatives labeled Winner and let AI recombine their proven elements — hook from one, body from another, offer from a third — into 5 remix briefs saved to a Canvas list.

Take your best-performing creators, embed their stories, posts and reels, and find which of your other creators produce the most similar content — so you know who to scale and where you overlap.

Every Monday: take the top active creatives, load their Creative DNA (hook, second-by-second timeline, transcript, claims, CTA) and write one row per creative into a Canvas list — one DNA field per column. Fully deterministic, no AI step: the DNA was captured once by the vision model at labeling time.

Weekly visual quality check of the top active creatives: is the product visible early, is every text overlay legible, is there a clear CTA? Answers from the cached Creative DNA when available; only un-analyzed assets get a live vision look. One combined report per run.